Second, the optionality of the rules. Something being presented as opt-in (you must specifically choose to use it), rather than opt-out (that's the rule, but you can elect to change the rules), makes it more acceptable to be presenting options that might perform better in one campaign than in another, because the intent is that only those campaigns that are improved (from the participants' view) by including an option will choose to include that option.
So that a combat-heavy campaign makes that -5/+10 look very appealing is the design working as intended, not something being "broken."
To me that's backwards thinking.
You don't include feats into a combat-heavy campaign because you want everybody to take those -5+10 feats.
If you include feats, no matter what focus it has, you do so because you hope that opt-in feature increases choice and variation, yes?
Well, if you agree to that I gotta say those feats mean design not working as intended, because in a combat-heavy campaign those feats are too good compared to the alternatives (other feats, no feats).
Actually, when I think about it, I have a hard time coming up with any style of campaign improved by opting for feats with your criteria in mind.
That is because I do not consider a campaign to be "improved" by adding a set of options where there are a very small subset which is clearly Best in Class for your style of campaign.
In contrast, my ideal opt-in rules module such as feats, would include half a dozen equally attractive feats for the combat pillar, half a dozen equally attractive feats for the explore pillar, and half a dozen equally attractive feats for the social pillar.
By your definition of design as intended, Aaron, I would say there ARE several feats working well, but that the -5+10 feats are not among them.